Flash won video before video was the question

Almost all the video you watch on the web arrives inside a plugin made by a graphics software company, and it got there without anyone in the industry deciding it should.

For most of a decade, watching something online meant choosing a player. There was Real’s, there was Microsoft’s, there was QuickTime, and each site picked one and told you to install it. So you installed all three, they each added something to your startup, they each wanted to be the default for everything, and roughly a third of the time you clicked a link and got a broken plugin icon.

Flash won that fight before video was even the question. It was already on virtually every machine because of banner advertising and browser games, and by the time anyone thought seriously about putting video on a web page, the install base was simply there.

That is the whole explanation. It is not a better codec. Early Flash video looked noticeably worse than what Real or Windows Media could manage at the same bitrate. It won because the user did not have to do anything, and one click beats a superior picture every time.

Adobe has said H.264 support is coming, which will close the quality gap and make the position close to unassailable.

Now consider where that leaves things. An enormous share of the moving pictures on the open web depends on a closed runtime from a single company, one that is free but not open, whose bugs you wait for them to fix, whose performance on anything other than Windows has always been an afterthought, and which decides on its own schedule which processors and operating systems are worth supporting.

Nobody chose this. It accumulated.

The reason I am thinking about it now is the phone Apple shipped in June, which has a genuinely capable browser and no Flash at all. Visit a video site on it and you get the icon, the same broken-plugin icon we spent ten years escaping, except this time there is no plugin to install because Apple has not permitted one.

Today that is a curiosity affecting a small number of expensive handsets. If devices of that kind become a meaningful share of how people reach the web, it stops being a curiosity and becomes a structural problem, because the runtime that carries the web’s video will be absent from a growing fraction of the machines viewing it.

Either Adobe gets onto those devices, which requires permission from a company with no reason to grant it, or the web needs a way to play video that does not involve a plugin at all. There is no third option, and the second one does not currently exist.